Source code for symfit.core.support

# SPDX-FileCopyrightText: 2014-2020 Martin Roelfs
#
# SPDX-License-Identifier: MIT

"""
This module contains support functions and convenience methods used
throughout symfit. Some are used predominantly internally, others are
designed for users.
"""
from __future__ import print_function
from collections import OrderedDict
import warnings
import re
import keyword

import numpy as np
from sympy.utilities.lambdify import lambdify
import sympy

from sympy.tensor import Idx
from sympy import symbols, MatrixExpr
from sympy.core.expr import Expr

from symfit.core.argument import Parameter, Variable
import symfit.core.printing  # Overwrites some numpy printing

import inspect
from functools import wraps


[docs]class deprecated(object): """ Decorator to raise a DeprecationWarning. """
[docs] def __init__(self, replacement=None): """ :param replacement: The function which should now be used instead. """ self.replacement = replacement
[docs] def __call__(self, func): @wraps(func) def deprecated_func(*args, **kwargs): warnings.warn(DeprecationWarning( '`{}` has been deprecated.'.format(func.__name__) + ' Use `{}` instead.'.format(self.replacement)) if self.replacement else '' ) return func(*args, **kwargs) return deprecated_func
[docs]def seperate_symbols(func): """ Seperate the symbols in symbolic function func. Return them in alphabetical order. :param func: scipy symbolic function. :return: (vars, params), a tuple of all variables and parameters, each sorted in alphabetical order. :raises TypeError: only symfit Variable and Parameter are allowed, not sympy Symbols. """ params = [] vars = [] for symbol in func.free_symbols: if not str(symbol).isidentifier(): continue # E.g. Indexed objects might print to A[i, j] if isinstance(symbol, Parameter): params.append(symbol) elif isinstance(symbol, Idx): # Idx objects are not seen as parameters or vars. pass elif isinstance(symbol, (MatrixExpr, Expr)): vars.append(symbol) else: raise TypeError('model contains an unknown symbol type, {}'.format(type(symbol))) for der in func.atoms(sympy.Derivative): # Used by jacobians and hessians, where derivatives are treated as # Variables. This way of writing it is purposefully discriminatory # against derivatives wrt variables, since such derivatives should be # performed explicitly in the case of jacs/hess, and are treated # differently in the case of ODEModels. if der.expr in vars and all(isinstance(s, Parameter) for s in der.variables): vars.append(der) params.sort(key=lambda symbol: symbol.name) vars.sort(key=lambda symbol: symbol.name) return vars, params
[docs]def sympy_to_py(func, args): """ Turn a symbolic expression into a Python lambda function, which has the names of the variables and parameters as it's argument names. :param func: sympy expression :param args: variables and parameters in this model :return: lambda function to be used for numerical evaluation of the model. """ # replace the derivatives with printable variables. derivatives = {var: Variable(var.name) for var in args if isinstance(var, sympy.Derivative)} func = func.xreplace(derivatives) args = [derivatives[var] if isinstance(var, sympy.Derivative) else var for var in args] lambdafunc = lambdify(args, func, dummify=False) # Check if the names of the lambda function are what we expect signature = inspect.signature(lambdafunc) sig_parameters = OrderedDict(signature.parameters) for arg, lambda_arg in zip(args, sig_parameters): if arg.name != lambda_arg: break else: # Lambdifying succesful! return lambdafunc # If we are here (very rare), then one of the lambda arg is still a Dummy. # In this case we will manually handle the naming. lambda_names = sig_parameters.keys() arg_names = [arg.name for arg in args] conversion = dict(zip(arg_names, lambda_names)) # Wrap the lambda such that arg names are translated into the correct dummy # symbol names @wraps(lambdafunc) def wrapped_lambdafunc(*ordered_args, **kwargs): converted_kwargs = {conversion[k]: v for k, v in kwargs.items()} return lambdafunc(*ordered_args, **converted_kwargs) # Update the signature of wrapped_lambdafunc to math our args new_sig_parameters = OrderedDict() for arg_name, dummy_name in conversion.items(): if arg_name == dummy_name: # Already has the correct name new_sig_parameters[arg_name] = sig_parameters[arg_name] else: # Change the dummy inspect.Parameter to the correct name param = sig_parameters[dummy_name] param = param.replace(name=arg_name) new_sig_parameters[arg_name] = param wrapped_lambdafunc.__signature__ = signature.replace( parameters=new_sig_parameters.values() ) return wrapped_lambdafunc
[docs]def sympy_to_scipy(func, vars, params): """ Convert a symbolic expression to one scipy digs. Not used by ``symfit`` any more. :param func: sympy expression :param vars: variables :param params: parameters :return: Scipy-style function to be used for numerical evaluation of the model. """ lambda_func = sympy_to_py(func, vars, params) def f(x, p): """ Scipy style function. :param x: list of arrays, NxM :param p: tuple of parameter values. """ x = np.atleast_2d(x) y = [x[i] for i in range(len(x))] if len(x[0]) else [] try: ans = lambda_func(*(y + list(p))) except TypeError: # Possibly this is a constant function in which case it only has Parameters. ans = lambda_func(*list(p))# * np.ones(x_shape) return ans return f
[docs]def variables(names, **kwargs): """ Convenience function for the creation of multiple variables. For more control, consider using ``symbols(names, cls=Variable, **kwargs)`` directly. :param names: string of variable names. Example: x, y = variables('x, y') :param kwargs: kwargs to be passed onto :func:`sympy.core.symbol.symbols` :return: iterable of :class:`symfit.core.argument.Variable` objects """ return symbols(names, cls=Variable, seq=True, **kwargs)
[docs]def parameters(names, **kwargs): """ Convenience function for the creation of multiple parameters. For more control, consider using ``symbols(names, cls=Parameter, **kwargs)`` directly. The `Parameter` attributes `value`, `min`, `max` and `fixed` can also be provided directly. If given as a single value, the same value will be set for all `Parameter`'s. When a sequence, it must be of the same length as the number of parameters created. Example:: x1, x2 = parameters('x1, x2', value=[2.0, 1.3], min=0.0) :param names: string of parameter names. Example: a, b = parameters('a, b') :param kwargs: kwargs to be passed onto :func:`sympy.core.symbol.symbols`. `value`, `min` and `max` will be handled separately if they are sequences. :return: iterable of :class:`symfit.core.argument.Parameter` objects """ sequence_fields = ['value', 'min', 'max', 'fixed'] sequences = {} for attr in sequence_fields: try: iter(kwargs[attr]) except (TypeError, KeyError): # Not iterable or not provided pass else: sequences[attr] = kwargs.pop(attr) if 'min' in sequences and 'max' in sequences: for min, max in zip(sequences['min'], sequences['max']): if min > max: raise ValueError('The value of `min` should be less than or' ' equal to the value of `max`.') params = symbols(names, cls=Parameter, seq=True, **kwargs) for key, values in sequences.items(): try: assert len(values) == len(params) except AssertionError: raise ValueError( '`len` of keyword-argument `{}` does not match the number of ' '`Parameter`s created.'.format(attr) ) except TypeError: # Iterator do not have a `len` but are allowed. pass finally: for param, value in zip(params, values): setattr(param, key, value) return params
[docs]class cached_property(property): """ A property which cashes the output of the first ever call and always returns that value from then on, unless delete is called on the attribute. This is typically used in converting `sympy` code into `scipy` compatible code, which is computationally a very expensive step we would like to perform only once. Does not allow setting of the attribute. """ base_str = '_cached'
[docs] def __init__(self, *args, **kwargs): super(cached_property, self).__init__(*args, **kwargs) self.cache_attr = '{}_{}'.format(self.base_str, self.fget.__name__)
[docs] def __get__(self, obj, objtype=None): """ In case of a first call, this will call the decorated function and return it's output. On every subsequent call, the same output will be returned. :param obj: the parent object this property is attached to. :param objtype: :return: Output of the first call to the decorated function. """ try: return getattr(obj, self.cache_attr) except AttributeError: # Call the wrapped function with the obj instance as argument setattr(obj, self.cache_attr, self.fget(obj)) return getattr(obj, self.cache_attr)
[docs] def __delete__(self, obj): """ Calling delete on the attribute will delete the cache. :param obj: parent object. """ try: delattr(obj, self.cache_attr) except AttributeError: pass
[docs]def jacobian(expr, symbols): """ Derive a symbolic expr w.r.t. each symbol in symbols. This returns a symbolic jacobian vector. :param expr: A sympy Expr. :param symbols: The symbols w.r.t. which to derive. """ jac = [] for symbol in symbols: # Differentiate to every param f = sympy.diff(expr, symbol) jac.append(f) return jac
[docs]def key2str(target): """ In ``symfit`` there are many dicts with symbol: value pairs. These can not be used immediately as \*\*kwargs, even though this would make a lot of sense from the context. This function wraps such dict to make them usable as \*\*kwargs immediately. :param target: `Mapping` to be made save :return: `Mapping` of str(symbol): value pairs. """ return target.__class__((str(symbol), value) for symbol, value in target.items())
def D(*args, **kwargs): # ToDo: Investigate sympy's inheritance properly to see if I can give a # .name atribute to Derivative objects or subclasses. return sympy.Derivative(*args, **kwargs)
[docs]def name(self): """ Save name which can be used for alphabetic sorting and can be turned into a kwarg. """ base_str = 'd{}{}_'.format(self.derivative_count if self.derivative_count > 1 else '', self.expr) for var, count in self.variable_count: base_str += 'd{}{}'.format(var, count if count > 1 else '') return base_str
sympy.Derivative.name = property(name)